INDEX

Explanations

phrases related to guidelines, criteria, or qualifications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

multimulti-layered wiring board

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 gyhoeddwyd

-0.56

RTEE

-0.51

RefNanny

-0.51

 ویکی‌پدی

-0.50

 transfieras

-0.46

MigrationBuilder

-0.46

trick

-0.44

MessageOf

-0.44

 Trick

-0.44

kript

-0.43

POSITIVE LOGITS

UrlResolution

0.44

addContainerGap

0.42

AndEndTag

0.41

 voetballer

0.37

 femei

0.36

 GoogleFonts

0.36

 medalla

0.36

 oamen

0.35

 relâche

0.34

 autorytatywna

0.34

Activations Density 0.002%

phrases related to guidelines, criteria, or qualifications

multimulti-layered wiring board

No Comments

No Known Activations